All training pictures are from the internet.
Do not use this model in commercials!
所有数据集从互联网收集,请不要用于非法用途,仅供学习使用。
Stable Diffusion model: chilloutmix_NiPrunedFp32Fix
<lora:evelyn:(recommend0.4-0.8 here)>,
Trigger word: evelyn
Sampling steps: 15~30
Sampling: Euler, DMP++ SDE Karras
when this model use to generate full-body photo, you need check the Hires.fix and set
step: 10~20
denoising strength: 0.4~0.7
upscale: 1.75
It will be work better.
v15: In this version, the half-body and full-body photos are better represented.
希望大家多多反馈意见和美图。